The sympathetic system reacts to affective states, resulting in electrodermal activity (EDA) changes. According to prior research, EDA could be associated with Bipolar Disorder (BD) mood episodes, but results remain inconclusive due to heterogeneous study design and the presence of cofounders. During sleep main confounders such as physical activity and marked temperature changes decrease. Therefore, we measured EDA with wearable devices during the day and night to explore whether EDA is a better biomarker of mood episodes during sleep. We analysed 140 participants from the TIMEBASE study, 102 with BD, and 28 healthy controls (HC). BD patients were in euthymic, manic, or depressive states. Participants were monitored with an E4 wearable device for 48 hours and symptoms were assessed through clinical scales. Overall, 15 EDA features were extracted from preprocessed data, and each was analysed with a mixed-effect model for repeated measures. Sleep state significantly influenced all EDA feature models except for phasic skewness. Similarly, its interaction with the group variable was significant in all models. During wakefulness, EDA features, especially phasic parameters, better distinguished between HC and types of BD episodes. Last, mixed symptoms, average temperature, and medication with anticholinergic effects were significant cofounders on most models, while sex barely influenced any model and age, none of them. We recommend using EDA collected from awake recordings to distinguish between BD groups, contrary to our initial hypothesis. These results approach research to daily monitoring of BD patients through wearables for episode prediction or treatment monitoring.